Hostname: page-component-7c8c6479df-7qhmt Total loading time: 0 Render date: 2024-03-28T17:24:39.329Z Has data issue: false hasContentIssue false

Immune factors and fatty acid composition in human milk from river/lake, coastal and inland regions of China

Published online by Cambridge University Press:  13 November 2012

Heidi J. Urwin
Affiliation:
Hugh Sinclair Unit of Human Nutrition, Department of Food and Nutritional Sciences, Whiteknights, PO Box 226, The University of Reading, ReadingRG6 6AP, UK
Jian Zhang
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China National Institute of Nutrition and Seafood Research (NIFES), N-5817Bergen, Norway Department of Biomedicine, University of Bergen, N-5009Bergen, Norway
Yixiong Gao
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China
Chunrong Wang
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China
Lixiang Li
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China
Pengkun Song
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China
Qingqing Man
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China
Liping Meng
Affiliation:
Institute of Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing100050, People's Republic of China
Livar Frøyland
Affiliation:
National Institute of Nutrition and Seafood Research (NIFES), N-5817Bergen, Norway
Elizabeth A. Miles
Affiliation:
The Human Development and Health Academic Unit, Faculty of Medicine, University of Southampton, Southampton, UK
Philip C. Calder
Affiliation:
The Human Development and Health Academic Unit, Faculty of Medicine, University of Southampton, Southampton, UK
Parveen Yaqoob*
Affiliation:
Hugh Sinclair Unit of Human Nutrition, Department of Food and Nutritional Sciences, Whiteknights, PO Box 226, The University of Reading, ReadingRG6 6AP, UK
*
*Corresponding author: Professor P. Yaqoob, email p.yaqoob@rdg.ac.uk
Rights & Permissions [Opens in a new window]

Abstract

Breast milk fatty acid composition may be affected by the maternal diet during gestation and lactation. The influence of dietary and breast milk fatty acids on breast milk immune factors is poorly defined. We determined the fatty acid composition and immune factor concentrations of breast milk from women residing in river/lake, coastal and inland regions of China, which differ in their consumption of lean fish and oily fish. Breast milk samples were collected on days 3–5 (colostrum), 14 and 28 post-partum (PP) and analysed for soluble CD14 (sCD14), transforming growth factor (TGF)-β1, TGF-β2, secretory IgA (sIgA) and fatty acids. The fatty acid composition of breast milk differed between the regions and with time PP. The concentrations of all four immune factors in breast milk decreased over time, with sCD14, sIgA and TGF-β1 being highest in the colostrum in the river and lake region. Breast milk DHA and arachidonic acid (AA) were positively associated, and γ-linolenic acid and EPA negatively associated, with the concentrations of each of the four immune factors. In conclusion, breast milk fatty acids and immune factors differ between the regions in China characterised by different patterns of fish consumption and change during the course of lactation. A higher breast milk DHA and AA concentration is associated with higher concentrations of immune factors in breast milk, suggesting a role for these fatty acids in promoting gastrointestinal and immune maturation of the infant.

Type
Full Papers
Copyright
Copyright © The Authors 2012 

The WHO(1) promotes exclusive breast-feeding for the first 6 months of life. Along with energy and nutrients, breast milk provides a broad range of immunological factors(Reference Chirico, Marzollo and Cortinovis2, Reference Walker3) of benefit to an infant during a time of gastrointestinal and immune system maturation. The long-chain PUFA (LC-PUFA) arachidonic acid (AA) and DHA are required for the development of the neonate(Reference Agostoni4). Fatty acids in breast milk can be affected by the maternal diet during both gestation(Reference Sauerwald, Demmelmair and Fidler5) and lactation(Reference Fidler, Sauerwald and Pohl6). LC n-3 PUFA are associated with reduced production of pro-inflammatory mediators and enhanced production of anti-inflammatory mediators(Reference Calder7). They may also modulate components of both natural (innate) and acquired (specific or adaptive) immunity(Reference Calder, Yaqoob and Thies8). There is evidence that exposure to LC-PUFA can influence aspects of immune function in early life(Reference Prescott and Dunstan9) and can have effects on later life(Reference Hanson and Gluckman10). However, the relationship between breast milk fatty acids and immune factors in breast milk is poorly defined.

In China, despite national food-based dietary guidelines(Reference Ge11), there are regional differences in diet due to tradition, local accessibility to food types and socio-economics(Reference Zhai, Wang and Du12). Regional differences in lean fish and oily fish intake provide an excellent opportunity to investigate dietary influences on breast milk composition. In the present study, we investigated differences in breast milk fatty acids and immune factors and the relationships between them. Previous studies in China have demonstrated significant differences in the fatty acid composition of breast milk from coastal, pastoral, rural and urban regions, where the populations differ in their dietary intakes of meat, fish, cereal and dairy products(Reference Li, Fan and Zhang13Reference Ruan, Liu and Man15), and differences in the macro- and micronutrient composition of breast milk between urban and rural areas(Reference Qian, Chen and Lu16). However, to our knowledge, there are no studies examining both breast milk fatty acid composition and immune factor concentrations in regions that differ in diet. Here, we report fatty acid composition and immune factor concentrations in breast milk collected at three time points from women living in three regions of China characterised by different patterns of fish consumption (lean v. oily and marine v. freshwater); women in the coastal region consume lean fish and some oily fish, while those in the river and lake region consume lean fish and little oily fish, and those in the inland region consume little in the way of fish. Furthermore, women in the coastal region consume mainly marine water fish that has a higher content of LC n-3 PUFA for an equivalent fat content compared with freshwater fish(Reference Du, Zhang and Wang17), which is mainly consumed by the women in the river and lake region. In a previous study, differences in the dietary intake of several fatty acids, including AA, EPA and DHA, and of LC-PUFA were seen between pregnant women living in coastal, river/lake and inland regions(Reference Zhang, Wang and Meng18).

Study design and methods used

Study design and subject characteristics

Aquamax China is a cross-sectional observational study of healthy pregnant women from three geographical regions of China associated with different dietary intakes of lean fish and oily fish and therefore of fatty acids. The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures were approved by the Ethics Committee of the Institute of Nutrition & Food Safety, Chinese Center for Disease Control and Prevention. All women gave written informed consent. Women with a healthy singleton pregnancy were recruited from hospitals in Jurong county, Jiangusu province (river and lake region, n 42), Rizhao county, Shandong province (coastal region, n 42) and Xushui county, Hebei province (inland region, n 41) between 18 and 22 weeks of gestation. Breast milk samples were collected during the first month post-partum (PP) to represent colostrum (days 3–5 PP), transition milk (day 14 PP) and mature milk (day 28 PP). The milk was collected between 09.00 and 10.00 hours before the baby had been fed (i.e. foremilk) and was manually expressed into sterile containers and immediately stored at − 20°C until transfer to the laboratory, where it was then stored at − 80°C until analysis.

Assessment of maternal diet

A semi-quantitative FFQ including 213 foods was developed(Reference Zhao, Hasegawa and Chen19, Reference Li, Song and Pan20) and used to estimate energy, macronutrient, fat and fatty acid intakes at 34 weeks of gestation. The FFQ was administered by interview by trained nutritionists or nurses and with the aid of food models, bowls, plates and cups. Nutrient intakes were calculated based on the Chinese Food Composition Tables(21).

Breast milk fatty acid composition

Lipids were extracted from whole milk (100 μl) using chloroform–methanol (2:1, v/v; 5 ml) as published previously(Reference Folch, Lees and Sloane Stanley22) with butylated hydroxytoluene (50 mg/l) added as an antioxidant. The mixture was emulsified with 1 m-NaCl (1 ml) and centrifuged at 715 g for 10 min. The lower organic layer was collected and dried under N2 at 40°C. Fatty acid methyl esters were prepared by the addition of toluene (400 μl) and 1·5 % H2SO4 in methanol (800 μl) followed by incubation for 1 h at 70°C. The mixture was neutralised with 0·1 m-K2CO3:0·1 m-KHCO3 (2 ml) and fatty acid methyl esters extracted into hexane (2 ml) and collected by centrifugation at 1120 g for 10 min; the upper layer was collected and dried under N2 at 40°C. Hexane (150 μl) was added to redissolve the dried sample, which was then stored at − 20°C until it was analysed by GC. Fatty acid methyl esters were resolved on a CP-SIL88 fused silica capillary column (50 m × 250 μm × 0·2 μm; Varian BV) using a HP6890 gas chromatograph (Agilent) equipped with a flame ionisation detector. Each sample (1 μl) was injected onto the column using an automatic injector at a split ratio of 100:1. He gas was used as a carrier. The detector and injector ports were set at 240 and 250°C, respectively. The oven temperature was maintained at 100°C for 4 min and the temperature was programmed to ramp to a maximum of 240°C at 3·5°C/min. Fatty acids of carbon lengths of 14–22 were identified by comparing retention times against authentic individual and mixed fatty acid methyl ester standards (NuChek Prep; Supelco Analytical). Fatty acids as a percentage of total fatty acids were determined from baseline corrected peak areas. Finally, three test samples of breast milk were analysed in duplicate for each new series of milk samples, and the average intra-assay CV was 1·5 % and the inter-assay CV was 4·1 %.

Breast milk immune factors

Immune factors were measured in whole milk, as described elsewhere(Reference Urwin, Miles and Noakes23). Intra- and inter-assay CV were determined from four replicates within a single assay and over a total of five separate assays, respectively. The concentrations of soluble CD14 (sCD14), transforming growth factor (TGF)-β1 and TGF-β2 were measured using the Quantikine Human sCD14, TGF-β1 and TGF-β2 immunoassays (R&D Systems Europe). The manufacturer's instructions were followed and whole-milk samples were assayed at dilutions of 10 000 for sCD14, 1·4 for TGF-β1 and 7·0 for TGF-β2. Latent TGF-β in breast milk was activated to the immunoreactive form as described for cell-culture supernatants with 1·0 m-HCl for 10 min and neutralised with 1·2 M-NaOH/0·5M-HEPES. For sCD14, the limit of detection was 125 pg/ml, the intra-assay CV 4 % and the inter-assay 11 %. For TGF-β1, the limit of detection was 7·0 pg/ml, the intra-assay CV 5 % and the inter-assay CV 15 %. For TGF-β2, the limit of detection was 7·0 pg/ml, the intra-assay CV 4 % and the inter-assay CV 9 %.

The concentration of sIgA was determined using the sIgA ELISA kit reported to have 100 % cross-reactivity for human sIgA from breast milk (Demeditec Diagnostics). The manufacturer's instructions were followed but with some adjustment for dilution factor. Whole-milk samples were assayed at a dilution of 15 000, and the limit of detection was 1·2 μg/ml, the intra-assay CV 7 % and the inter-assay CV 16 %.

All plates were read on a GENios Spectra FLUOR plus (Tecan UK Limited).

Breast milk protein content

The summary of a workshop on immune factors in human milk(Reference Filteau24) describes the difficulties in standardising samples when working in the field and recommended measuring immune factors both per unit volume and standardised against another milk component, e.g. K or total protein. Thus, the protein content of breast milk was determined using a standard method based on that of Bradford(Reference Bradford25, Reference Bergqvist, Karisson and Fohlin26). Human serum albumin (Sigma-Aldrich) was diluted with distilled water to produce a range of protein standards from 0·125 to 2·0 mg/ml. Standards or diluted whole-breast milk samples (10 μl) were mixed with Bradford reagent (500 μl). Immediately, 200 μl aliquots were transferred to wells of a ninety-six-well flat-bottom plate, and absorbance read at 595 nm on a GENios Spectra FLUOR plus. The limit of detection was 0·1 mg/ml, the intra-assay CV was 7·0 % and the inter-assay CV was 9·0 %.

Statistics

Data were checked for normality with histograms and the Kolmogorov–Smirnov test. Data are expressed as mean and standard deviation or median and 25th-75th percentiles as appropriate. Subject characteristics were analysed with the Kruskal–Wallis test, Fisher's exact test or χ2 test as appropriate. Breast milk fatty acids were calculated as a percentage of total fatty acids. Breast milk immune factors were calculated as either absolute concentration or concentration per mg protein. All data were transformed into ranks and a rank ANCOVA for the effect of time, region and region × time interaction were determined, adjusting for covariates (parity, maternal age, maternal BMI and self-reported maternal atopy). Where appropriate, post hoc tests with Bonferroni correction were determined. Partial correlations of breast milk immune factors with dietary fatty acids were carried out on rank transformed data, controlling for region and repeated measures. Models for predicting breast milk immune factors were developed using multiple regression analysis. The independent variables selected for inclusion in the model were determined from initial significant bivariate correlations with the dependent variable. All multiple regression analyses were applied in conditions that ensured a suitable fit. These conditions were explored using relevant residual analyses and plots. The Statistical Package for Social Sciences version 15 (SPSS, Inc.) was used for all tests, and in all cases, a P value of < 0·05 was considered as statistically significant (P< 0·05 adjusted for multiple comparisons).

Results

Subject characteristics

The characteristics of the pregnant women from the three regions are shown in Table 1. Women from the coastal region were significantly older than women from the other two regions, and, in addition, there was a greater proportion of multipara women in the coastal region. Weight and BMI were different between the river/lake and coastal regions, with women from the coastal region weighing more and having a higher BMI. The weights and BMI of women from the inland region were not significantly different from the other two regions. The coastal region had the highest proportion of women self-reporting atopic disease. Overall, the level of education in the inland region was the poorest, with a majority of women having only attended junior high school. The number of women with a university or higher education was significantly higher in the river and lake region. All women were of Han ethnicity.

Table 1 Characteristics of the pregnant women across the three geographical regions of China (Numbers of subjects and within-group percentages; medians and 25th–75th percentiles)

a,bValues within a row with unlike superscript letters were significantly different (P< 0·05; Kruskall–Wallis test).

* Groups compared by Fisher's exact test or χ2 test as appropriate.

Post hoc tests, based on the standardised residual being lower or higher than the critical values − 1·96,+1·96, respectively, to be significant at P< 0·05.

Self-reported eczema, allergic rhinitis, asthma and atopic dermatitis.

Maternal diet

Intake of lean fish by women from the river/lake and the coastal regions did not differ, but women from both of these regions had a significantly higher intake of lean fish than those from the inland region (Table 2). Women from the coastal region had a higher intake of oily fish compared with those from the river and lake region (Table 2).

Table 2 Daily dietary characteristics of the pregnant women at 34 weeks of pregnancy across the three geographical regions of China (Mean values and standard deviations; medians and 25th–75th percentiles)

% TE, percentage of total energy.

a,b,cValues within a row with unlike superscript letters were significantly different (P< 0·05).

* Groups compared by Kruskall–Wallis test.

Groups compared by ANOVA.

Energy intakes were comparable among pregnant women in the three regions (Table 2). Women from the river and lake region consumed less protein as a percentage of total energy compared with those from the coastal region, and less fat and more carbohydrate as percentage of total energy compared with those from the inland region (Table 2). Women from the river and lake region had the highest intake of total n-3 PUFA, but this was mostly contributed by α-linolenic acid (ALA) (Table 2). In contrast, women from the coastal region had the lowest intake of total n-3 PUFA, and the highest dietary ratio of total n-6:total n-3 PUFA (Table 2). However, the intakes of EPA, docosapentaenoic acid (DPA) and DHA were significantly higher for women from the coastal region compared with those in the other two regions (Table 2). Women from the inland region had the highest intake of total n-6 PUFA, including linoleic acid (LA) and AA, but the lowest intake of EPA and DPA, while the intake of DHA was comparable with that of the river and lake region (Table 2).

Breast milk fatty acid composition

The fatty acid composition of breast milk from the three regions of China is shown in Table 3, where the data are presented as a percentage of total breast milk fatty acids. There was a significant effect of region and/or time and/or a significant region × time interaction for the majority of fatty acids (Table 3).

Table 3 Breast milk fatty acid composition (individual fatty acids as a percentage of total fatty acids) at days 3 to 5, 14 and 28 post-partum (PP) across the three geographical regions of China (Medians and 25th–75th percentiles)

R, region; T, time; LA, linoleic acid; AA, arachidonic acid; LC, long chain; ALA, α-linolenic acid; DPA, docosapentaenoic acid.

a,b,c Median values within a row with unlike superscript letters were significantly different (P< 0·05).

* P values for two-factor rank ANCOVA adjusting for maternal age, BMI, parity and atopy.

Median values were significantly different from the day 3–5 sample (colostrum) (P< 0·05).

Median values were significantly different from the day 14 sample (transition milk) (P< 0·05).

SFA

There was a significant effect of region (P< 0·001) and time (P= 0·001) and a region × time interaction (P< 0·001) for the proportion of total SFA (Table 3). These were lowest in the colostrum from the river and lake region (P< 0·05) and highest in the transition and mature milk from the inland region (both P< 0·05). In the river and lake region and the inland region, the percentage of total SFA increased between the colostrum and transition milk, whereas it decreased in the coastal region (Table 3).

MUFA

There was a significant effect of region and time (P< 0·001 for both) for the proportion of total MUFA; however, the region × time interaction was not significant (P= 0·121) (Table 3). MUFA were higher in breast milk from the river and lake region compared with the coastal and inland regions at all time points (P< 0·05 for all) and higher in the coastal region compared with the inland region in the colostrum and transition milk (P< 0·05 for both). In the river and lake region, the percentage of oleic acid and of total MUFA decreased with time. However, in the coastal and inland regions, these did not differ significantly over time (Table 3).

PUFA

There was a significant effect of region (P< 0·001) and time (P= 0·007) and a region × time interaction (P< 0·001) for total n-6 PUFA (Table 3). The percentage of LA was lowest in the milk from the river and lake region at all three time points (Table 3). The proportion of LA in the milk from the coastal and inland regions was not significantly different, except for the colostrum, where LA was lower in the coastal region (P< 0·05) (Table 3). LA increased with time in the river/lake and coastal regions, whereas it did not differ significantly with time in the inland region. There was a significant region × time interaction (P< 0·001) for the total LC n-6 PUFA; in transition milk, they were lower in the river and lake region compared with the coastal region, but the inland region did not differ from either. The total LC n-6 PUFA decreased between the colostrum and transition milk from the river and lake region and the inland region and between transition milk and mature milk from the coastal region (Table 3). The proportion of AA did not differ between the regions in the colostrum or mature milk, but was lower in the transition milk from the river and lake region compared with the coastal region; the inland region did not differ from either.

There was a significant effect of region and time and a region × time interaction for total n-3 PUFA and total LC n-3 PUFA (P< 0·001 for all). The percentage of ALA was highest in the milk from the river and lake region at all three time points (P< 0·05 for all), whereas it was comparable in the coastal and inland regions, except for mature milk, where it was higher in the coastal region. EPA, DPA and DHA were also higher in the river and lake region compared with the inland region at all three time points (P< 0·05 for all). Compared with the coastal region, the river and lake region had higher total LC n-3 PUFA in the colostrum (P< 0·05), a comparable content in transition milk and a lower content in mature milk (P< 0·05). The total LC n-3 PUFA in the river/lake and inland regions decreased with time, which was not seen in the coastal region.

The total n-6:total n-3 PUFA ratio and the total n-6:total n-3 LC-PUFA ratio were significantly different between the three regions (Table 3): both ratios were highest in the inland region and lowest in the river and lake region at all three time points (P< 0·05 for all) with the exception of mature milk, when the total n-6:total n-3 LC-PUFA ratio in the river/lake and coastal regions was not significantly different.

The relationships between dietary fatty acids and breast milk fatty acids are shown in Table 4. There were positive correlations between dietary LC n-3 PUFA and breast milk LC n-3 PUFA; however, there was no correlation between dietary AA and breast milk AA (Table 4). There was a positive correlation between dietary ALA and breast milk ALA and EPA, but no correlation with DPA and DHA. Dietary ALA was negatively correlated with milk LA and, conversely, dietary LA was positively correlated with milk LA and negatively correlated with milk LC n-3 PUFA and ALA (Table 4).

Table 4 Relationship between dietary fatty acid intakes (g/d) and breast milk fatty acids (% of total fatty acids)

LA, linoleic acid; AA, arachidonic acid; ALA, α-linolenic acid; DPA, docosapentaenoic acid.

P< 0·05, ** P< 0·01, *** P< 0·001.

Dietary intakes of fatty acids are from FFQ administered at 34 weeks of pregnancy.

Values are partial correlation coefficients of rank-transformed data, for the whole population, controlling for repeated measures.

Breast milk protein

There was a significant effect of region (P= 0·009) and time (P< 0·001) and a region × time interaction (P= 0·001) for breast milk total protein (Table 5). The protein concentration was highest in the colostrum from the river and lake region, and it was comparable across all three regions in the transition milk and lowest in the mature milk from the coastal region (Table 5). There was no significant effect of parity, maternal age, BMI or maternal self-reported atopy on breast milk total protein (data not shown).

Table 5 Breast milk protein and immune factor concentrations at days 3 to 5, 14 and 28 post-partum (PP) across the three geographical regions of China (Medians and 25th–75th percentiles)

R, region; T, time; sCD14, soluble CD14; sIgA, secretory IgA; TGF-β, transforming growth factor-β.

a,b,cMedian values within a row with unlike superscript letters were significantly different (P< 0·05).

* P values for two-factor rank ANCOVA adjusting for maternal age, BMI, parity and atopy.

Median values were significantly different from the day 3–5 sample (colostrum) (P< 0·05).

Median values were significantly different from the day 14 sample (transition milk) (P< 0·05).

Breast milk immune factors

There was a 100 % detection rate for each of the four immune factors measured. The concentrations of the breast milk immune factors (sCD14, sIgA, TGF-β1 and TGF-β2) for each region are shown in Table 5. Unadjusted analysis (Kruskall–Wallis) identified significant differences between the regions for all immune factors at at least one time point (data not shown). A more detailed analysis using rank-transformed data in an ANCOVA (rank ANCOVA), which considered maternal age, BMI, parity and self-reported atopy as covariates, identified a significant effect of time and a significant region × time interaction for each of the immune factors and also a significant region effect for TGF-β1 (P= 0·043; Table 5). The concentrations of sCD14, sIgA and TGF-β1 were higher in the colostrum from the river and lake region compared with the coastal and inland regions (mean 1·5-, 3·0- and 1·4-fold higher, respectively; P= 0·05 for all) and did not differ between the coastal and inland regions. Colostrum TGF-β2 was also higher in the river and lake region compared with the coastal region (1·3-fold) and the inland region (1·7-fold); however, in each case, this was not statistically significant. In transition milk, sCD14 and sIgA were not significantly different across the regions, but TGF-β1 in the coastal region was 2·2-fold higher than that in the river and lake region (P= 0·05) and 1·5 fold higher than that in the inland region (NS). TGF-β2 concentration in the transition milk from the coastal region was 1·9-fold higher than that from the river and lake region (P= 0·05) and 1·4-fold higher than that from the inland region (NS). In mature milk, only TGF-β1 differed significantly between the regions, again being higher in the coastal region compared with the river/lake (1·6-fold, P= 0·05) and the inland regions (1·3-fold, NS); the river and lake region and the inland region were not significantly different from each other. For all regions, each of the immune factors decreased with time (Table 5). When the data were adjusted for total protein content (concentration per mg protein), a significant region × time interaction remained for all immune factors (P< 0·01 for all; data not shown), with the exception of sCD14, where there was a trend towards significance (P= 0·077; data not shown). There was also a significant effect of time for each of the immune factors (P< 0·01 for all; data not shown), an effect of region for TGF-β1 (P= 0·001; data not shown) and a trend towards an effect of region for TGF-β2 (P= 0·078; data not shown).

There were positive partial correlations (controlling for region and repeated measures) among all the immune factors: TGF-β1 with TGF-β2 (r 0·681), sIgA (r 0·541) and sCD14 (r 0·536); TGF-β2 with sIgA (r 0·443) and sCD14 (r 0·450); sIgA with sCD14 (r 0·564) (P< 0·001 for all).

There were no significant effects of maternal age, BMI and parity on breast milk immune factors. There was a significant maternal atopy × time interaction for TGF-β1 and TGF-β2; both were higher in the transition milk from atopic women compared with non-atopic women (P< 0·05 for both; data not shown). However, when region was used as a covariate in an ANCOVA, the effect of maternal atopy on TGF-β1 was no longer significant.

The influence of individual variables in multiple linear regression analyses of the breast milk immune factors is shown in Table 6. In both models, maternal age, BMI, self-reported atopy and parity were included but their effects are not shown in the table; maternal self-reported atopy was not a significant predictor for any of the immune factors measured. Maternal age showed a trend towards being a weak positive predictor of TGF-β1 (β = 0·100, P< 0·075) and TGF-β2 (β = 0·123, P< 0·05). Parity had a trend towards being a weak negative predictor for sCD14 (β = 0·090, P< 0·075) and TGF-β2 (β = 0·114, P< 0·060), while BMI was a significant but weak positive predictor of sIgA (β = 0·080, P< 0·05) and TGF-β1 (β = 0·114, P< 0·01). In general, the change from colostrum to transition milk predicted a decrease in the concentration of each of the immune factors and the change from transition to mature milk predicted a further decrease in sCD14 and sIgA (Table 6). Breast milk γ-linolenic acid was a negative predictor and breast milk AA a positive predictor of all immune factors, except for TGF-β2, where AA was not a significant predictor (Table 6). Due to concerns of multi-collinearity, breast milk n-3 PUFA were analysed in a separate model and DHA was a positive predictor of all immune factors and EPA a negative predictor for almost all, with a trend towards being a negative predictor for TGF-β2 (Table 6). When AA and DHA were replaced by dihomo-γ-linolenic acid and DPA in the two separate models, respectively, again to avoid multi-collinearity, both were positive predictors of each of the immune factors measured (P< 0·001 for all, data not shown) except for TGF-β2, where milk dihomo-γ-linolenic acid was not a significant predictor.

Table 6 Influence of individual variables in multiple linear regression models for the breast milk immune factors* (β Coefficients and P values)

sCD14, soluble CD14; sIgA, secretory IgA; TGF-β, transforming growth factor-β; PP, post-partum.

* The characteristics of the mother (age, BMI, self-reported atopy and parity) were included in each of the models but the results are not shown.

β Standardised regression coefficient.

There were no significant partial correlations (controlling for region and time) between dietary stearic acid, total SFA, oleic acid and total MUFA during pregnancy with any of the breast milk immune factors (data not shown). The partial correlations between milk immune factors and dietary intakes of n-6 and n-3 fatty acids during pregnancy are shown in Table 7. Dietary intakes of EPA, DPA and DHA were positively correlated with TGF-β1 and TGF-β2, and there was a negative correlation of dietary ALA intake with TGF-β1 (Table 7). Dietary intakes of LA and AA did not correlate with any of the immune factors (Table 7).

Table 7 Relationships between the breast milk immune factors and dietary fatty acid intakes (g/d) during pregnancy

LA, linoleic acid; ALA, α-linolenic acid; AA, arachidonic acid; DPA, docosapentaenoic acid; sCD14, soluble CD14; sIgA, secretory IgA; TGF-β, transforming growth factor-β.

P< 0·05, ** P< 0·01, *** P< 0·001.

Dietary intakes of fatty acids are from FFQ administered at 34 weeks of pregnancy.

Values are partial correlation coefficients of rank-transformed data, controlling for region and time.

Discussion

Women in the three geographical regions studied here have different patterns of consumption of several foods, including freshwater/marine and lean/oily fish. Women in the coastal region consume lean and some oily fish, while those in the river and lake region consume lean fish and those in the inland region consume little in the way of fish. Consequently, pregnant women in these three regions have different intakes of EPA, DHA and total LC n-3 PUFA, as reported previously for pregnant Chinese women living in such regions(Reference Zhang, Wang and Meng18). Previous studies have reported differences in breast milk fatty acids in Chinese women living in urban, rural, pastoral, coastal and inland regions, which are largely attributable to diet(Reference Li, Fan and Zhang13Reference Ruan, Liu and Man15). The present study confirms significant differences in the fatty acid composition of breast milk from regions of China differing in their consumption of aquatic foods. However, for the first time, it demonstrates differences in the concentrations of breast milk immune factors between these regions and, further, that consumption of fatty acids by the women during pregnancy and the fatty acid composition of breast milk are important predictors of breast milk immune factor concentrations. Breast milk immune factors play a key role in the defence of the newborn infant(Reference Chirico, Marzollo and Cortinovis2) and in their early gastrointestinal and immune maturation(Reference Walker3). Thus, these observations suggest an important link between maternal diet, breast milk fatty acids, breast milk immune factors and infant development.

Previous reports have shown that breast milk lipid content varies significantly between mothers(Reference Prentice, Prentice and Whitehead27) and increases between the start and the end of a feed (i.e. between foremilk and hindmilk)(Reference Prentice, Prentice and Whitehead27, Reference Neville, Keller and Seacat28), and that the difference in breast milk lipid content between the end of one feed and the start of another is inversely related to the duration of time between the feeds(Reference Prentice, Prentice and Whitehead27). However, despite reported differences in breast milk lipid content between foremilk and hindmilk, there was no significant difference in the breast milk fatty acid composition between foremilk and hindmilk(Reference Gibson and Kneebone29). In the present study, breast milk samples were collected before the baby had been fed (i.e. foremilk) and they were analysed for breast milk fatty acid composition and not breast milk lipid content. The consumption of an oily fish meal during lactation has a transient acute effect on breast milk LC n-3 PUFA composition, and the time lapse between eating the oily fish meal and the sampling of breast milk influences the size of the effect observed(Reference Lauritzen, Jorgensen and Hansen30). In the present study, breast milk samples were collected at the same time each day between 09.00 and 10.00 hours. It has previously been reported that breast milk protein does not vary significantly between foremilk and hindmilk but does vary significantly between mothers(Reference Khan, Casadio and Lai31).

In the present study, the proportion of LC n-3 PUFA in the colostrum was highest in the river and lake>coastal>inland region and the proportion of milk DHA was of the order river and lake = coastal>inland region, even though women in the coastal region had the highest intake of preformed dietary LC n-3 PUFA. However, women in the river and lake region had the highest intake of total n-3 PUFA, mainly due to ALA, compared with the coastal and inland regions. Fatty acids in breast milk can be derived from the diet(Reference Fidler, Sauerwald and Pohl6), from adipose tissue stores(Reference Sauerwald, Demmelmair and Fidler5) or from endogenous synthesis(Reference Burdge and Wootton32). Although conversion of ALA to DHA in humans is limited, it has been shown to be enhanced in women taking the contraceptive pill containing 17α-ethynyloestradiol(Reference Burdge and Wootton32). Thus, it is possible that the higher circulating oestrogen levels during pregnancy enhance the conversion of ALA to DHA and so increase the content of DHA in adipose tissue. Indeed, studies in rodents have shown that consuming a diet rich in ALA from conception results in DHA accretion rates that are similar to those achievable with dietary DHA(Reference Valenzuela, Von Bernhardi and Valenzuela33, Reference Childs, Romijn and Enke34) and that fetal DHA deficiency can be largely reversed with maternal ALA consumption(Reference Harauma, Salem and Moriguchi35). However, a study of pregnant women supplemented with ALA and LA did not report a significant increase in maternal and neonatal DHA status, only in EPA and DPA(Reference de Groot, Hornstra and van Houwelingen36). This may have been due to the fact that women were only supplemented during the second and third trimesters with 2·8 g ALA/d, whereas women from the river and lake region in the present study had a high habitual intake of ALA (mean 4·6 g/d). Interestingly, total LC n-3 PUFA in milk from the river/lake and inland regions decreased with time, whereas women from the coastal region had the highest dietary intake of preformed LC n-3 PUFA, and milk proportions of LC n-3 PUFA did not change significantly over time, so that in mature milk, the levels in the coastal region were the highest (coastal>river and lake>inland).

Milk production is stimulated by prolactin, which suppresses oestrogen activity, which in turn could down-regulate the previously enhanced conversion of ALA to DHA. Thus, it is plausible that the endogenous conversion of ALA to LC n-3 PUFA is enhanced under the influence of oestrogen during pregnancy, but not maintained during lactation, when oestrogen is suppressed. Indeed, lactating women who were supplemented for 4 weeks with flax seed oil (providing 10·7 g ALA/d) did not experience an increase in milk DHA(Reference Francois, Connor and Bolewicz37). The ratio of n-6:n-3 PUFA in the diet may also provide a plausible explanation for the differences in the proportions of LC n-3 PUFA in breast milk among the three regions. Although women in the coastal region had the highest intake of preformed LC n-3 PUFA, the dietary n-6:n-3 PUFA ratio was 15:1 compared with 5:1 in the river and lake region and 10:1 in the inland region. Diets high in LA are associated with a reduced incorporation of preformed LC n-3 PUFA into tissue phospholipids in non-human primates(Reference McMurchie, Rinaldi and Burnard38), and decreasing the LA content of infant formula is as effective as increasing the ALA content with regard to increasing EPA and DHA in infant erythrocytes(Reference Clark, Makrides and Neumann39). In the present study, there were weak, but significant, negative correlations between dietary LA and both breast milk ALA and LC n-3 PUFA, and positive correlations between dietary ALA and breast milk ALA and EPA.

The concentrations of each of the immune factors decreased within the first month of lactation, as previously reported by others(Reference Castellote, Casillas and Ramirez-Santana40); this coincides with an increase in the volume of breast milk consumed by the infant over the same period of time(Reference Neville, Keller and Seacat41). Between days 3 and 28 PP, milk intake by the infant approximately doubles(Reference Burdge and Wootton32). In the present study, the concentrations of the milk immune factors typically decreased by 50–60 % between these times (although sIgA decreased by 90 % in the river and lake region), suggesting that, in general, the actual intake of the immune factors (except sIgA in the river and lake region) would not be altered. There were also significant regional differences in immune factors, which were influenced by the stage of lactation. Colostrum from the river and lake region, which had the highest proportion of LC n-3 PUFA, had the highest concentrations of sCD14, sIgA and TGF-β1; the concentration of TGF-β2 was also higher, but this was not statistically significant. There was a positive association of milk TGF-β1 with milk DPA, DHA, dihomo-γ-linolenic acid and AA, which agrees with the findings of Ferrucci et al. (Reference Ferrucci, Cherubini and Bandinelli42), where there was a positive correlation of TGF-β with DHA and AA in human plasma. However, in a previous intervention study with oily fish(Reference Urwin, Miles and Noakes23) and in a high-dose fish oil intervention study(Reference Hawkes, Bryan and Neumann43), increasing dietary LC n-3 PUFA during the latter half of pregnancy did not significantly affect the concentrations of TGF-β1 or TGF-β2 in breast milk, and there was no association of milk DHA with milk TGF-β. This suggests that factors other than dietary fatty acid intake and/or breast milk fatty acid composition play a role in the regional differences in breast milk immune factors.

Breast milk sCD14 was higher in the colostrum from women from the river and lake region. There was a positive association between sCD14 and breast milk AA and a negative association with milk γ-linolenic acid, as previously reported by others in maternal serum and breast milk, respectively(Reference Laitinen, Hoppu and Hamalainen44). Dunstan et al. (Reference Dunstan, Roper and Mitoulas45) were the first to report a positive correlation between sCD14 and milk DPA (but not DHA) in the colostrum after a high-dose fish oil intervention. In the present study, there was a positive association between sCD14 and DPA during the first month of lactation, and sCD14 was positively associated with milk DHA and negatively with milk EPA. The lack of a correlation between sCD14 and milk DHA and EPA content in the study of Dunstan et al. (Reference Dunstan, Roper and Mitoulas45) may possibly be explained by the fact that they measured sCD14 only in the colostrum, whereas we report sCD14 concentrations at three time points throughout the first month of lactation. Also, the DHA content of the colostrum in the Dunstan et al. study(Reference Dunstan, Roper and Mitoulas45) was much higher at >1·0 % of fatty acids, which coincided with a concomitant decrease in AA.

Breast milk sIgA was positively associated with AA and negatively with EPA. Some studies have reported that the AA-derived mediator, PGE2, promotes IgA synthesis(Reference Garrone, Galibert and Rousset46, Reference Pestka47) and since LC n-3 PUFA may reduce PGE2, this could explain the negative association with EPA. However, there was also a positive association of sIgA with DPA and DHA in both the present study and that of Dunstan et al. (Reference Dunstan, Roper and Mitoulas45), and this may be linked to the positive association between DPA, DHA and TGF-β, the latter of which is involved in the switching of B cells to produce sIgA(Reference Suzuki and Fagarasan48).

TGF-β is one of the most abundant cytokines in human milk(Reference Kalliomaki, Ouwehand and Arvilommi49). It has pleiotropic effects, including promoting oral tolerance(Reference Ando, Hatsushika and Wako50), cell proliferation and differentiation(Reference Clark and Coker51, Reference Sonoda, Matsumoto and Hitoshi52) and protection against allergy-related outcomes in infancy and childhood(Reference Oddy and Rosales53). There have been contrasting reports on the effect of maternal atopy on concentrations of TGF-β in breast milk(Reference Urwin, Miles and Noakes23, Reference Oddy and Rosales53). In the present study, TGF-β1 and TGF-β2 were higher in the transition milk of atopic women compared with non-atopic women. However, the majority of atopic women lived in the coastal region and when this was considered in the analysis, the effect of atopy was no longer significant.

sCD14 provides the neonate with protection against pathogenic invasion, since the CD14-independent pathway in response to endotoxin is lacking in neonates(Reference Cohen, Haziot and Shen54). Several studies have associated higher concentrations of sCD14 in breast milk with a protective effect towards infant and childhood atopy(Reference Jones, Holloway and Popplewell55Reference Savilahti, Siltanen and Kajosaari57), although this is not unequivocal(Reference Snijders, Damoiseaux and Penders58). Also, newborns do not produce their own protective levels of sIgA until 30 d PP, but breast-fed neonates benefit from the supply and passive protection afforded by sIgA within their mothers' milk(Reference Xanthou, Bines and Walker59). Reductions in these immune factors could potentially alter the maturation of the neonatal immune system, the risk of allergy(Reference Oddy and Rosales53, Reference Laiho, Lampi and Hamalainen60) and the overall health(Reference Hanson61) of the infant.

The present cross-sectional study took advantage of dietary contrasts in three geographical regions, and while it is unable to deduce causality, it provides valuable insight into the associations between breast milk immune factors and dietary and breast milk fatty acids. It is likely that factors other than dietary and breast milk fatty acids play a role in regional differences in breast milk composition; many women in China still follow traditional cultural practices during the first month PP(Reference Liu, Mao and Sun62), with restrictions on diet, hygiene and physical activity, but available details regarding these practices are limited.

In conclusion, breast milk fatty acids and immune factors differ between the regions in China characterised by different patterns of lean fish and oily fish consumption and the change during the course of lactation. A higher breast milk DHA and AA concentration is associated with higher concentrations of immune factors in breast milk, suggesting a role for these fatty acids in promoting gastrointestinal and immune maturation of the infant.

Acknowledgements

This study was supported by the European Commission under Framework 6: sustainable aquafeeds to maximise the health benefits of farmed fish for consumers (Aquamax: FOOD-CT-2006-16249). We thank all mothers and their families for their contribution. We also thank the following staff from the three participating institutions: Ms Rong Yu and Ms Hong Shang from the Health Centre for pregnant women and infants in Jurong City, Jiangsu Province; Ms Pan Xiaoqun from Jiangsu CDC; Ms An Fengjuan, Ms Lan Wang and Mr Xianping Zhang from the Hospital of Rizhao Harbor City, Shangdong Province; Ms Mingxing Yang, Ms Jianying Hu and Mr Yantao Li from the Health Centre for pregnant women and infants in Xushui County, Hebei Province. Thanks also to Dr Caroline E. Childs, Hugh Sinclair Unit of Human Nutrition, University of Reading for guidance in breast milk fatty acid analysis, and to Professor Sue Todd, University of Reading for statistical advice. J. Z., L. F., E. A. M. and P. C. C. designed the study. Y. G., C. W., L. L., P. S., Q. M. and L. M. recruited the subjects, and collected and processed the samples under the supervision of J. Z. H. J. U. conducted the breast milk analysis under the supervision of P. Y. H. J. U. conducted the statistical analysis and wrote the first draft of the paper. All authors contributed to and approved the final version of the manuscript. There are no conflicts of interest to report.

References

1World Health Organization & UNICEF (2003) Global Strategy for Infant and Young Child Feeding. Geneva: WHO.Google Scholar
2Chirico, G, Marzollo, R, Cortinovis, S, et al. (2008) Antiinfective properties of human milk. J Nutr 138, 1801S1806S.Google Scholar
3Walker, A (2010) Breast milk as the gold standard for protective nutrients. J Pediatr 156, S3S7.Google Scholar
4Agostoni, C (2008) Role of long-chain polyunsaturated fatty acids in the first year of life. J Pediatr Gastroenterol Nutr 47, Suppl. 2, S41S44.CrossRefGoogle ScholarPubMed
5Sauerwald, TU, Demmelmair, H, Fidler, N, et al. (2000) Polyunsaturated fatty acid supply with human milk. Physiological aspects and in vivo studies of metabolism. Adv Exp Med Biol 478, 261270.Google Scholar
6Fidler, N, Sauerwald, T, Pohl, A, et al. (2000) Docosahexaenoic acid transfer into human milk after dietary supplementation: a randomized clinical trial. J Lipid Res 41, 13761383.Google Scholar
7Calder, PC (2006) n-3 Polyunsaturated fatty acids, inflammation, and inflammatory diseases. Am J Clin Nutr 83, 1505S1519S.Google Scholar
8Calder, PC, Yaqoob, P, Thies, F, et al. (2002) Fatty acids and lymphocyte functions. Br J Nutr 87, Suppl. 1, S31S48.CrossRefGoogle ScholarPubMed
9Prescott, SL & Dunstan, JA (2007) Prenatal fatty acid status and immune development: the pathways and the evidence. Lipids 42, 801810.Google Scholar
10Hanson, M & Gluckman, P (2011) Developmental origins of noncommunicable disease: population and public health implications. Am J Clin Nutr 94, 1754S1758S.Google Scholar
11Ge, K (2011) The transition of Chinese dietary guidelines and food guide pagoda. Asia Pac J Clin Nutr 20, 439446.Google Scholar
12Zhai, F, Wang, H, Du, S, et al. (2007) Lifespan nutrition and changing socio-economic conditions in China. Asia Pac J Clin Nutr 16, Suppl.1, 374382.Google Scholar
13Li, J, Fan, Y, Zhang, Z, et al. (2009) Evaluating the trans fatty acid, CLA, PUFA and erucic acid diversity in human milk from five regions in China. Lipids 44, 257271.Google Scholar
14Peng, Y, Zhou, T, Wang, Q, et al. (2009) Fatty acid composition of diet, cord blood and breast milk in Chinese mothers with different dietary habits. Prostaglandins Leukot Essent Fatty Acids 81, 325330.Google Scholar
15Ruan, C, Liu, X, Man, H, et al. (1995) Milk composition in women from five different regions of China: the great diversity of milk fatty acids. J Nutr 125, 29932998.Google Scholar
16Qian, J, Chen, T, Lu, W, et al. (2010) Breast milk macro- and micronutrient composition in lactating mothers from suburban and urban Shanghai. J Paediatr Child Health 46, 115120.Google Scholar
17Du, YZ, Zhang, J, Wang, C, et al. (2012) Risk-benefit evaluation of fish from Chinese markets: nutrients and contaminants in 24 fish species from five big cities and related assessment for human health. Sci Total Environ 416, 187199.CrossRefGoogle ScholarPubMed
18Zhang, J, Wang, Y, Meng, L, et al. (2009) Maternal and neonatal plasma n-3 and n-6 fatty acids of pregnant women and neonates in three regions in China with contrasting dietary patterns. Asia Pac J Clin Nutr 18, 377388.Google Scholar
19Zhao, W, Hasegawa, K & Chen, J (2002) The use of food-frequency questionnaires for various purposes in China. Public Health Nutr 5, 829833.CrossRefGoogle ScholarPubMed
20Li, YP, Song, J, Pan, H, et al. (2006) Validity of food frequency questionnaire to investigate the dietary energy and nutrients intake. Acta Nutr Sin 28, 143147.CrossRefGoogle Scholar
21Institute of Nutrition and Food Safety, Chinese Center for Disease Control (2002 and 2005) Chinese Food Composition 2002 and Chinese Food Composition 2004. Beijing: Peking University Medical Press.Google Scholar
22Folch, J, Lees, M & Sloane Stanley, GH (1957) A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem 226, 497509.Google Scholar
23Urwin, HJ, Miles, EA, Noakes, PS, et al. (2012) Salmon consumption during pregnancy alters fatty acid composition and secretory IgA concentration in human breast milk. J Nutr 142, 16031610.CrossRefGoogle ScholarPubMed
24Filteau, S (2009) Measuring trace immune factors in human milk: workshop summary. Adv Exp Med Biol 639, 331337.Google Scholar
25Bradford, MM (1976) A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72, 248254.Google Scholar
26Bergqvist, Y, Karisson, L & Fohlin, L (1989) Total protein determined in human breast milk by use of Coomassie brilliant blue and centrifugal analysis. Clin Chem 35, 21272129.Google Scholar
27Prentice, A, Prentice, AM & Whitehead, RG (1981) Breast-milk fat concentrations of rural African women. 1. Short-term variations within individuals. Br J Nutr 45, 483494.Google Scholar
28Neville, MC, Keller, RP, Seacat, J, et al. (1984) Studies on human lactation. I. Within-feed and between-breast variation in selected components of human milk. Am J Clin Nutr 40, 635646.CrossRefGoogle ScholarPubMed
29Gibson, RA & Kneebone, GM (1980) Effect of sampling on fatty acid composition of human colostrum. J Nutr 110, 16711675.Google Scholar
30Lauritzen, L, Jorgensen, MH, Hansen, HS, et al. (2002) Fluctuations in human milk long-chain PUFA levels in relation to dietary fish intake. Lipids 37, 237244.CrossRefGoogle ScholarPubMed
31Khan, S, Casadio, YS, Lai, CT, et al. (2012) Investigation of short-term variations in casein and whey proteins in breastmilk of term mothers. J Pediatr Gastroenterol Nutr 55, 136141.Google Scholar
32Burdge, GC & Wootton, SA (2002) Conversion of alpha-linolenic acid to eicosapentaenoic, docosapentaenoic and docosahexaenoic acids in young women. Br J Nutr 88, 411420.CrossRefGoogle ScholarPubMed
33Valenzuela, A, Von Bernhardi, R, Valenzuela, V, et al. (2004) Supplementation of female rats with alpha-linolenic acid or docosahexaenoic acid leads to the same omega-6/omega-3 LC-PUFA accretion in mother tissues and in fetal and newborn brains. Ann Nutr Metab 48, 2835.CrossRefGoogle ScholarPubMed
34Childs, CE, Romijn, T, Enke, U, et al. (2010) Maternal diet during pregnancy has tissue-specific effects upon fetal fatty acid composition and alters fetal immune parameters. Prostaglandins Leukot Essent Fatty Acids 83, 179184.Google Scholar
35Harauma, A, Salem, N Jr & Moriguchi, T (2010) Repletion of n-3 fatty acid deficient dams with alpha-linolenic acid: effects on fetal brain and liver fatty acid composition. Lipids 45, 659668.Google Scholar
36de Groot, RH, Hornstra, G, van Houwelingen, AC, et al. (2004) Effect of alpha-linolenic acid supplementation during pregnancy on maternal and neonatal polyunsaturated fatty acid status and pregnancy outcome. Am J Clin Nutr 79, 251260.Google Scholar
37Francois, CA, Connor, SL, Bolewicz, LC, et al. (2003) Supplementing lactating women with flaxseed oil does not increase docosahexaenoic acid in their milk. Am J Clin Nutr 77, 226233.CrossRefGoogle Scholar
38McMurchie, EJ, Rinaldi, JA, Burnard, SL, et al. (1990) Incorporation and effects of dietary eicosapentaenoate (20 : 5(n-3)) on plasma and erythrocyte lipids of the marmoset following dietary supplementation with differing levels of linoleic acid. Biochim Biophys Acta 1045, 164173.Google Scholar
39Clark, KJ, Makrides, M, Neumann, MA, et al. (1992) Determination of the optimal ratio of linoleic acid to alpha-linolenic acid in infant formulas. J Pediatr 120, S151S158.Google Scholar
40Castellote, C, Casillas, R, Ramirez-Santana, C, et al. (2011) Premature delivery influences the immunological composition of colostrum and transitional and mature human milk. J Nutr 141, 11811187.Google Scholar
41Neville, MC, Keller, R, Seacat, J, et al. (1988) Studies in human lactation: milk volumes in lactating women during the onset of lactation and full lactation. Am J Clin Nutr 48, 13751386.Google Scholar
42Ferrucci, L, Cherubini, A, Bandinelli, S, et al. (2006) Relationship of plasma polyunsaturated fatty acids to circulating inflammatory markers. J Clin Endocrinol Metab 91, 439446.CrossRefGoogle ScholarPubMed
43Hawkes, JS, Bryan, DL, Neumann, MA, et al. (2001) Transforming growth factor beta in human milk does not change in response to modest intakes of docosahexaenoic acid. Lipids 36, 11791181.Google Scholar
44Laitinen, K, Hoppu, U, Hamalainen, M, et al. (2006) Breast milk fatty acids may link innate and adaptive immune regulation: analysis of soluble CD14, prostaglandin E2, and fatty acids. Pediatr Res 59, 723727.CrossRefGoogle ScholarPubMed
45Dunstan, JA, Roper, J, Mitoulas, L, et al. (2004) The effect of supplementation with fish oil during pregnancy on breast milk immunoglobulin A, soluble CD14, cytokine levels and fatty acid composition. Clin Exp Allergy 34, 12371242.CrossRefGoogle ScholarPubMed
46Garrone, P, Galibert, L, Rousset, F, et al. (1994) Regulatory effects of prostaglandin E2 on the growth and differentiation of human B lymphocytes activated through their CD40 antigen. J Immunol 152, 42824290.Google Scholar
47Pestka, JJ (2003) Deoxynivalenol-induced IgA production and IgA nephropathy-aberrant mucosal immune response with systemic repercussions. Toxicol Lett 140–141, 287295.Google Scholar
48Suzuki, K & Fagarasan, S (2009) Diverse regulatory pathways for IgA synthesis in the gut. Mucosal Immunol 2, 468471.Google Scholar
49Kalliomaki, M, Ouwehand, A, Arvilommi, H, et al. (1999) Transforming growth factor-beta in breast milk: a potential regulator of atopic disease at an early age. J Allergy Clin Immunol 104, 12511257.Google Scholar
50Ando, T, Hatsushika, K, Wako, M, et al. (2007) Orally administered TGF-beta is biologically active in the intestinal mucosa and enhances oral tolerance. J Allergy Clin Immunol 120, 916923.CrossRefGoogle ScholarPubMed
51Clark, DA & Coker, R (1998) Transforming growth factor-beta (TGF-beta). Int J Biochem Cell Biol 30, 293298.Google Scholar
52Sonoda, E, Matsumoto, R, Hitoshi, Y, et al. (1989) Transforming growth factor beta induces IgA production and acts additively with interleukin 5 for IgA production. J Exp Med 170, 14151420.Google Scholar
53Oddy, WH & Rosales, F (2010) A systematic review of the importance of milk TGF-beta on immunological outcomes in the infant and young child. Pediatr Allergy Immunol 21, 4759.Google Scholar
54Cohen, L, Haziot, A, Shen, DR, et al. (1995) CD14-independent responses to LPS require a serum factor that is absent from neonates. J Immunol 155, 53375342.Google Scholar
55Jones, CA, Holloway, JA, Popplewell, EJ, et al. (2002) Reduced soluble CD14 levels in amniotic fluid and breast milk are associated with the subsequent development of atopy, eczema, or both. J Allergy Clin Immunol 109, 858866.CrossRefGoogle ScholarPubMed
56Rothenbacher, D, Weyermann, M, Beermann, C, et al. (2005) Breastfeeding, soluble CD14 concentration in breast milk and risk of atopic dermatitis and asthma in early childhood: birth cohort study. Clin Exp Allergy 35, 10141021.Google Scholar
57Savilahti, E, Siltanen, M, Kajosaari, M, et al. (2005) IgA antibodies, TGF-beta1 and -beta2, and soluble CD14 in the colostrum and development of atopy by age 4. Pediatr Res 58, 13001305.Google Scholar
58Snijders, BE, Damoiseaux, JG, Penders, J, et al. (2006) Cytokines and soluble CD14 in breast milk in relation with atopic manifestations in mother and infant (KOALA Study). Clin Exp Allergy 36, 16091615.Google Scholar
59Xanthou, M, Bines, J & Walker, WA (1995) Human milk and intestinal host defense in newborns: an update. Adv Pediatr 42, 171208.Google Scholar
60Laiho, K, Lampi, AM, Hamalainen, M, et al. (2003) Breast milk fatty acids, eicosanoids, and cytokines in mothers with and without allergic disease. Pediatr Res 53, 642647.Google Scholar
61Hanson, LA (2007) Session 1: feeding and infant development breast-feeding and immune function. Proc Nutr Soc 66, 384396.Google Scholar
62Liu, N, Mao, L, Sun, X, et al. (2006) Postpartum practices of puerperal women and their influencing factors in three regions of Hubei, China. BMC Public Health 6, 274.CrossRefGoogle ScholarPubMed
Figure 0

Table 1 Characteristics of the pregnant women across the three geographical regions of China (Numbers of subjects and within-group percentages; medians and 25th–75th percentiles)

Figure 1

Table 2 Daily dietary characteristics of the pregnant women at 34 weeks of pregnancy across the three geographical regions of China (Mean values and standard deviations; medians and 25th–75th percentiles)

Figure 2

Table 3 Breast milk fatty acid composition (individual fatty acids as a percentage of total fatty acids) at days 3 to 5, 14 and 28 post-partum (PP) across the three geographical regions of China (Medians and 25th–75th percentiles)

Figure 3

Table 4 Relationship between dietary fatty acid intakes (g/d)† and breast milk fatty acids (% of total fatty acids)‡

Figure 4

Table 5 Breast milk protein and immune factor concentrations at days 3 to 5, 14 and 28 post-partum (PP) across the three geographical regions of China (Medians and 25th–75th percentiles)

Figure 5

Table 6 Influence of individual variables in multiple linear regression models for the breast milk immune factors* (β Coefficients and P values)

Figure 6

Table 7 Relationships between the breast milk immune factors and dietary fatty acid intakes† (g/d) during pregnancy‡